linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Jim Cromie <jim.cromie@gmail.com>
To: linux-kernel@vger.kernel.org, jbaron@akamai.com,
	gregkh@linuxfoundation.org, ukaszb@chromium.org,
	louis.chauvet@bootlin.com
Cc: dri-devel@lists.freedesktop.org, amd-gfx@lists.freedesktop.org,
	intel-gvt-dev@lists.freedesktop.org,
	intel-gfx@lists.freedesktop.org, daniel.vetter@ffwll.ch,
	tvrtko.ursulin@linux.intel.com, jani.nikula@intel.com,
	ville.syrjala@linux.intel.com, seanpaul@chromium.org,
	robdclark@gmail.com, groeck@google.com, yanivt@google.com,
	bleung@google.com, quic_saipraka@quicinc.com, will@kernel.org,
	catalin.marinas@arm.com, quic_psodagud@quicinc.com,
	maz@kernel.org, arnd@arndb.de,
	linux-arm-kernel@lists.infradead.org,
	linux-arm-msm@vger.kernel.org, mingo@redhat.com,
	jim.cromie@gmail.com
Subject: [PATCH v4 28/58] dyndbg: restore classmap protection when theres a controlling_param
Date: Sat,  2 Aug 2025 21:57:46 -0600	[thread overview]
Message-ID: <20250803035816.603405-29-jim.cromie@gmail.com> (raw)
In-Reply-To: <20250803035816.603405-1-jim.cromie@gmail.com>

DRM has always had /sys/module/drm/parameters/debug (ie drm.debug).
Without dyndbg, this is their only control point.  One could presume
they like it - in any case its a system/user interface, ie ABI.

With dyndbg enabled, drm calls DYNAMIC_DEBUG_CLASSMAP_PARAM() to
create the drm.debug kparam, wired to our param-handler, which writes
a "class FOO" query for each bit in the classmap.  Since no new
interface was ever contemplated, this is using >control.

Since drm.debug is ABI, we should not allow class-less queries to
alter our implementation of its settings.

This patch provides that protection, *only* when theres a PARAM.  This
is the user, expressing their wish for easy control of their entire
classmap.  They also wish to trust its settings.

Classes without a PARAM are unprotected, allowing admins their
shortcuts.  No such use-cases exist yet.

Anyway, this patch does:

1. adds struct _ddebug_class_map.controlling_param

2. set it in ddebug_match_apply_kparam(), during modprobe/init,
   when options like drm.debug are handled.

3. ddebug_class_has_param() checks .controlling_param

4. ddebug_class_wants_protection() macro renames 3.

5. ddebug_change() calls 4 when needed.
   IE when applying a class-less query to a class'd pr_debug / drm_dbg_<T>

Historical Summary:

-v0 - before classmaps.  no special case keywords
-v1 - "class DEFAULT" is assumed if not mentioned.
      this protects classes from class-less queries

-v2.pre-this-patch - protection macro'd to false
-v2.with-this-patch - sysfs knob decides
-v2.speculative - module decides wrt classmap protection
		  seems unneeded now, TBD

NOTE: protection is only against class-less queries, explicit "class
FOO" adjustments are allowed (that is the mechanism).

Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
v3 - new patch
---
 include/linux/dynamic_debug.h | 16 ++++++----
 lib/dynamic_debug.c           | 55 ++++++++++++++++++++++++++---------
 2 files changed, 52 insertions(+), 19 deletions(-)

diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
index 2d959f1f8cd30..bc26bc9128c1c 100644
--- a/include/linux/dynamic_debug.h
+++ b/include/linux/dynamic_debug.h
@@ -100,8 +100,9 @@ enum ddebug_class_map_type {
  * __pr_debug_cls(0, "fake CORE msg") in any part of DRM would "work"
  * __pr_debug_cls(22, "no such class") would compile, but not "work"
  */
-
+struct _ddebug_class_param;
 struct _ddebug_class_map {
+	struct _ddebug_class_param *controlling_param;
 	const struct module *mod;		/* NULL for builtins */
 	const char *mod_name;
 	const char **class_names;
@@ -231,7 +232,12 @@ struct _ddebug_class_param {
  *
  * Creates a sysfs-param to control the classes defined by the
  * exported classmap, with bits 0..N-1 mapped to the classes named.
- * This version keeps class-state in a private long int.
+ *
+ * Since sysfs-params are ABI, this also protects the classmap'd
+ * pr_debugs from un-class'd `echo -p > /proc/dynamic_debug/control`
+ * changes.
+ *
+ * This keeps class-state in a private long int.
  */
 #define DYNAMIC_DEBUG_CLASSMAP_PARAM(_name, _var, _flags)		\
 	static unsigned long _name##_bvec;				\
@@ -244,10 +250,8 @@ struct _ddebug_class_param {
  * @_var:   name of the (exported) classmap var defining the classes/bits
  * @_flags: flags to be toggled, typically just 'p'
  *
- * Creates a sysfs-param to control the classes defined by the
- * exported clasmap, with bits 0..N-1 mapped to the classes named.
- * This version keeps class-state in user @_bits.  This lets drm check
- * __drm_debug elsewhere too.
+ * Like DYNAMIC_DEBUG_CLASSMAP_PARAM, but maintains param-state in
+ * extern @_bits.  This lets DRM check __drm_debug elsewhere too.
  */
 #define DYNAMIC_DEBUG_CLASSMAP_PARAM_REF(_name, _bits, _var, _flags)	\
 	__DYNAMIC_DEBUG_CLASSMAP_PARAM(_name, _bits, _var, _flags)
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index 548a82a178d49..c3e27637d9357 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -198,16 +198,26 @@ ddebug_find_valid_class(struct _ddebug_info const *di, const char *query_class,
 }
 
 /*
- * classmaps-v1 protected classes from changes by legacy commands
- * (those selecting _DPRINTK_CLASS_DFLT by omission), v2 undoes that
- * special treatment.  State so explicitly.  Later we could give
- * modules the choice to protect their classes or to keep v2 behavior.
+ * classmaps-V1 protected classes from changes by legacy commands
+ * (those selecting _DPRINTK_CLASS_DFLT by omission).  This had the
+ * downside that saying "class FOO" for every change can get tedious.
+ *
+ * V2 is smarter, it protects class-maps if the defining module also
+ * calls DYNAMIC_DEBUG_CLASSMAP_PARAM to create a sysfs parameter.
+ * Since they want the knob, we should assume they intend to use it
+ * (in preference to "class FOO +p" >control), and want to trust its
+ * settings.
+ * This gives protection when its useful, and not when its just tedious.
  */
-static inline bool ddebug_client_module_protects_classes(const struct ddebug_table *dt)
+static inline bool ddebug_class_has_param(const struct _ddebug_class_map *map)
 {
-	return false;
+	return !!(map->controlling_param);
 }
 
+/* re-framed as a policy choice */
+#define ddebug_class_wants_protection(map) \
+	ddebug_class_has_param(map)
+
 /*
  * Search the tables for _ddebug's which match the given `query' and
  * apply the `flags' and `mask' to them.  Returns number of matching
@@ -250,7 +260,7 @@ static int ddebug_change(const struct ddebug_query *query, struct flag_settings
 					/* site.class != given class */
 					continue;
 				/* legacy query, class'd site */
-				else if (ddebug_client_module_protects_classes(dt))
+				else if (ddebug_class_wants_protection(map))
 					continue;
 				/* allow change on class'd pr_debug */
 			}
@@ -650,6 +660,7 @@ static int ddebug_exec_queries(char *query, const char *modname)
 }
 
 /* apply a new class-param setting */
+
 static int ddebug_apply_class_bitmap(const struct _ddebug_class_param *dcp,
 				     const unsigned long *new_bits,
 				     const unsigned long old_bits,
@@ -1228,25 +1239,36 @@ static void ddebug_sync_classbits(const struct kernel_param *kp, const char *mod
 	}
 }
 
-static void ddebug_match_apply_kparam(const struct kernel_param *kp,
-				      const struct _ddebug_class_map *map,
-				      const char *mod_name)
+static struct _ddebug_class_param *
+ddebug_get_classmap_kparam(const struct kernel_param *kp,
+			   const struct _ddebug_class_map *map)
 {
 	struct _ddebug_class_param *dcp;
 
 	if (kp->ops != &param_ops_dyndbg_classes)
-		return;
+		return false;
 
 	dcp = (struct _ddebug_class_param *)kp->arg;
 
-	if (map == dcp->map) {
+	return (map == dcp->map)
+		? dcp : (struct _ddebug_class_param *)NULL;
+}
+
+static void ddebug_match_apply_kparam(const struct kernel_param *kp,
+				      struct _ddebug_class_map *map,
+				      const char *mod_name)
+{
+	struct _ddebug_class_param *dcp = ddebug_get_classmap_kparam(kp, map);
+
+	if (dcp) {
+		map->controlling_param = dcp;
 		v2pr_info(" kp:%s.%s =0x%lx", mod_name, kp->name, *dcp->bits);
 		vpr_cm_info(map, " %s mapped to: ", mod_name);
 		ddebug_sync_classbits(kp, mod_name);
 	}
 }
 
-static void ddebug_apply_params(const struct _ddebug_class_map *cm, const char *mod_name)
+static void ddebug_apply_params(struct _ddebug_class_map *cm, const char *mod_name)
 {
 	const struct kernel_param *kp;
 #if IS_ENABLED(CONFIG_MODULES)
@@ -1266,6 +1288,13 @@ static void ddebug_apply_params(const struct _ddebug_class_map *cm, const char *
 	}
 }
 
+/*
+ * called from add_module, ie early. it can find controlling kparams,
+ * which can/does? enable protection of this classmap from class-less
+ * queries, on the grounds that the user created the kparam, means to
+ * use it, and expects it to reflect reality.  We should oblige him,
+ * and protect those classmaps from classless "-p" changes.
+ */
 static void ddebug_apply_class_maps(const struct _ddebug_info *di)
 {
 	struct _ddebug_class_map *cm;
-- 
2.50.1



  parent reply	other threads:[~2025-08-03  5:15 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-08-03  3:57 [PATCH v4 00/58] fix CONFIG_DRM_USE_DYNAMIC_DEBUG=y Jim Cromie
2025-08-03  3:57 ` [PATCH v4 01/58] docs/dyndbg: update examples \012 to \n Jim Cromie
2025-08-03  3:57 ` [PATCH v4 02/58] docs/dyndbg: explain flags parse 1st Jim Cromie
2025-08-03  3:57 ` [PATCH v4 03/58] test-dyndbg: fixup CLASSMAP usage error Jim Cromie
2025-08-03  3:57 ` [PATCH v4 04/58] dyndbg: reword "class unknown," to "class:_UNKNOWN_" Jim Cromie
2025-08-03  3:57 ` [PATCH v4 05/58] dyndbg: make ddebug_class_param union members same size Jim Cromie
2025-08-03  3:57 ` [PATCH v4 06/58] dyndbg: drop NUM_TYPE_ARRAY Jim Cromie
2025-08-03  3:57 ` [PATCH v4 07/58] dyndbg: tweak pr_fmt to avoid expansion conflicts Jim Cromie
2025-08-03  3:57 ` [PATCH v4 08/58] dyndbg: reduce verbose/debug clutter Jim Cromie
2025-08-03  3:57 ` [PATCH v4 09/58] dyndbg: refactor param_set_dyndbg_classes and below Jim Cromie
2025-08-03  3:57 ` [PATCH v4 10/58] dyndbg: tighten fn-sig of ddebug_apply_class_bitmap Jim Cromie
2025-08-03  3:57 ` [PATCH v4 11/58] dyndbg: replace classmap list with a vector Jim Cromie
2025-08-03  3:57 ` [PATCH v4 12/58] dyndbg: macrofy a 2-index for-loop pattern Jim Cromie
2025-08-03  3:57 ` [PATCH v4 13/58] dyndbg,module: make proper substructs in _ddebug_info Jim Cromie
2025-08-03  3:57 ` [PATCH v4 14/58] dyndbg: hoist classmap-filter-by-modname up to ddebug_add_module Jim Cromie
2025-08-03  3:57 ` [PATCH v4 15/58] dyndbg: ddebug_table.mod_name down to _ddebug_info Jim Cromie
2025-08-03  3:57 ` [PATCH v4 16/58] dyndbg-API: remove DD_CLASS_TYPE_(DISJOINT|LEVEL)_NAMES and code Jim Cromie
2025-08-03  3:57 ` [PATCH v4 17/58] selftests-dyndbg: add a dynamic_debug run_tests target Jim Cromie
2025-08-03  3:57 ` [PATCH v4 18/58] dyndbg: change __dynamic_func_call_cls* macros into expressions Jim Cromie
2025-08-03  3:57 ` [PATCH v4 19/58] dyndbg-API: replace DECLARE_DYNDBG_CLASSMAP Jim Cromie
2025-08-03  3:57 ` [PATCH v4 20/58] dyndbg: detect class_id reservation conflicts Jim Cromie
2025-08-03  3:57 ` [PATCH v4 21/58] dyndbg: check DYNAMIC_DEBUG_CLASSMAP_DEFINE args at compile-time Jim Cromie
2025-08-03  3:57 ` [PATCH v4 22/58] dyndbg-test: change do_prints testpoint to accept a loopct Jim Cromie
2025-08-03  3:57 ` [PATCH v4 23/58] dyndbg-API: promote DYNAMIC_DEBUG_CLASSMAP_PARAM to API Jim Cromie
2025-08-03  3:57 ` [PATCH v4 24/58] dyndbg: treat comma as a token separator Jim Cromie
2025-08-03  3:57 ` [PATCH v4 25/58] dyndbg: split multi-query strings with % Jim Cromie
2025-08-03  3:57 ` [PATCH v4 26/58] selftests-dyndbg: add test_mod_submod Jim Cromie
2025-08-03  3:57 ` [PATCH v4 27/58] dyndbg: drop "protection" of class'd pr_debugs from legacy queries Jim Cromie
2025-08-03  3:57 ` Jim Cromie [this message]
2025-08-03  3:57 ` [PATCH v4 29/58] docs/dyndbg: add classmap info to howto Jim Cromie
2025-08-03  3:57 ` [PATCH v4 30/58] dyndbg: reserve flag-bit _DPRINTK_FLAGS_PREFIX_CACHED Jim Cromie
2025-08-03  3:57 ` [PATCH v4 31/58] dyndbg: add _DPRINTK_FLAGS_INCL_LOOKUP for +mfsl flags Jim Cromie
2025-08-03  3:57 ` [PATCH v4 32/58] dyndbg: refactor *dynamic_emit_prefix to split lookups Jim Cromie
2025-08-03  3:57 ` [PATCH v4 33/58] drm: use correct ccflags-y spelling Jim Cromie
2025-08-03  3:57 ` [PATCH v4 34/58] drm-dyndbg: adapt drm core to use dyndbg classmaps-v2 Jim Cromie
2025-08-03  3:57 ` [PATCH v4 35/58] drm-dyndbg: adapt DRM to invoke DYNAMIC_DEBUG_CLASSMAP_PARAM Jim Cromie
2025-08-03  3:57 ` [PATCH v4 36/58] drm-print: fix config-dependent unused variable Jim Cromie
2025-08-03  3:57 ` [PATCH v4 37/58] drm-dyndbg: DRM_CLASSMAP_USE in amdgpu driver Jim Cromie
2025-08-03  3:57 ` [PATCH v4 38/58] drm-dyndbg: DRM_CLASSMAP_USE in i915 driver Jim Cromie
2025-08-03  3:57 ` [PATCH v4 39/58] drm-dyndbg: DRM_CLASSMAP_USE in drm_crtc_helper Jim Cromie
2025-08-03  3:57 ` [PATCH v4 40/58] drm-dyndbg: DRM_CLASSMAP_USE in drm_dp_helper Jim Cromie
2025-08-03  3:57 ` [PATCH v4 41/58] drm-dyndbg: DRM_CLASSMAP_USE in nouveau Jim Cromie
2025-08-03  3:58 ` [PATCH v4 42/58] drm-dyndbg: add DRM_CLASSMAP_USE to Xe driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 43/58] drm-dyndbg: add DRM_CLASSMAP_USE to virtio_gpu Jim Cromie
2025-08-03  3:58 ` [PATCH v4 44/58] drm-dyndbg: add DRM_CLASSMAP_USE to simpledrm Jim Cromie
2025-08-03  3:58 ` [PATCH v4 45/58] drm-dyndbg: add DRM_CLASSMAP_USE to bochs Jim Cromie
2025-08-03  3:58 ` [PATCH v4 46/58] drm-dyndbg: add DRM_CLASSMAP_USE to etnaviv Jim Cromie
2025-08-03  3:58 ` [PATCH v4 47/58] drm-dyndbg: add DRM_CLASSMAP_USE to gma500 driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 48/58] drm-dyndbg: add DRM_CLASSMAP_USE to radeon Jim Cromie
2025-08-03  3:58 ` [PATCH v4 49/58] drm-dyndbg: add DRM_CLASSMAP_USE to vmwgfx driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 50/58] drm-dyndbg: add DRM_CLASSMAP_USE to vkms driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 51/58] drm-dyndbg: add DRM_CLASSMAP_USE to udl driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 52/58] drm-dyndbg: add DRM_CLASSMAP_USE to mgag200 driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 53/58] drm-dyndbg: add DRM_CLASSMAP_USE to the gud driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 54/58] drm-dyndbg: add DRM_CLASSMAP_USE to the qxl driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 55/58] drm-dyndbg: add DRM_CLASSMAP_USE to the drm_gem_shmem_helper driver Jim Cromie
2025-08-03  3:58 ` [PATCH v4 56/58] drm: restore CONFIG_DRM_USE_DYNAMIC_DEBUG un-BROKEN Jim Cromie
2025-08-03  3:58 ` [PATCH v4 57/58] amdxdna: add -DDYNAMIC_DEBUG_MODULE to cflags - RFC Jim Cromie
2025-08-03  3:58 ` [PATCH v4 58/58] accel: add -DDYNAMIC_DEBUG_MODULE to subdir-cflags " Jim Cromie

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250803035816.603405-29-jim.cromie@gmail.com \
    --to=jim.cromie@gmail.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=arnd@arndb.de \
    --cc=bleung@google.com \
    --cc=catalin.marinas@arm.com \
    --cc=daniel.vetter@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=groeck@google.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-gvt-dev@lists.freedesktop.org \
    --cc=jani.nikula@intel.com \
    --cc=jbaron@akamai.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=louis.chauvet@bootlin.com \
    --cc=maz@kernel.org \
    --cc=mingo@redhat.com \
    --cc=quic_psodagud@quicinc.com \
    --cc=quic_saipraka@quicinc.com \
    --cc=robdclark@gmail.com \
    --cc=seanpaul@chromium.org \
    --cc=tvrtko.ursulin@linux.intel.com \
    --cc=ukaszb@chromium.org \
    --cc=ville.syrjala@linux.intel.com \
    --cc=will@kernel.org \
    --cc=yanivt@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).